Insert Vals Into SQL From Access That Don't Exist

I've got an access table with about 2 million rows. I'm using this to update a table in SQL that holds pretty much the exact same data, only with an added Identity column.

From week to week, the access table grows. For example, next week it may have 2.1 millions rows, the week after 2.2 million, etc.

The goal of the DTS is to keep the SQL table up to date w/ the access one. In the past, this has been done by deleting everything from the SQL Table and then importing the ENTIRE access table. This not only takes more time then need be, since the majority of the records *already* existed, but it also threw referential integrity out the door - other tables should be referencing the Identity in the SQL Table. IDEALLY, the only rows that would be transferred from the access file are ones that don't already exist in the SQL table.

I don't want to re-invent the wheel, and have to confess being a little under-schooled on all that SSIS has to offer. Is there a Data Flow Transformation that would solve this?? Any other advice? If all else fails, I'd probably just dump the entire access table to a temp table and then insert vals into the production table that don't exist, but even this would require more temp hard drive space then I'd like.

Thanks!